sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling buggrafted

alb_break_llc() decides whether to break LLC preference during active
load balance. It does so by testing that every runnable fair task on the
source rq prefers its LLC:

	env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable

But the two counters cover different sets. nr_pref_llc_running is updated
in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued,
so it follows queued tasks. h_nr_runnable is updated in set_delayed()/
clear_delayed() and drops delay-dequeued tasks.

So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted
in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks,
alb_break_llc() returns false, and active balance is free to pull a task
off its preferred LLC. Active balance only moves runnable tasks, and this
is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB
skips the per-task test in can_migrate_task(). The runnable set is the one
we want.

Fix it on the counter side. A task should be counted in
nr_pref_llc_running exactly while it is both queued on its preferred LLC
(pref_llc_queued) and runnable (!sched_delayed). Define that membership
once in task_pref_llc_runnable(), and adjust the counter only through
pref_llc_running_inc()/pref_llc_running_dec() from the four sites that
change either input: account_llc_enqueue(), account_llc_dequeue(),
set_delayed() and clear_delayed(). Gating every update on the same
predicate keeps the delay, wake and dequeue paths from double-counting
or underflowing; see the comments at those sites for the ordering.

nr_llc_running and sd->llc_counts are not touched and stay on queued
semantics.

Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC")
Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/
Reported-by: Zhan Xusheng <zhanxusheng@xiaomi.com>
Suggested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Cc: <stable@kernel.org> # v7.2.x
Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com

1 files changed, 114 insertions, 0 deletions

diff --git a/include/drm/drm_suballoc.h b/include/drm/drm_suballoc.h
new file mode 100644
index 000000000..29befdda3
--- /dev/null
+++ b/include/drm/drm_suballoc.h
@@ -0,0 +1,114 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Copyright 2011 Red Hat Inc.
+ * Copyright © 2022 Intel Corporation
+ */
+#ifndef _DRM_SUBALLOC_H_
+#define _DRM_SUBALLOC_H_
+
+#include <drm/drm_mm.h>
+
+#include <linux/dma-fence.h>
+#include <linux/types.h>
+
+#define DRM_SUBALLOC_MAX_QUEUES 32
+/**
+ * struct drm_suballoc_manager - fenced range allocations
+ * @wq: Wait queue for sleeping allocations on contention.
+ * @hole: Pointer to first hole node.
+ * @olist: List of allocated ranges.
+ * @flist: Array[fence context hash] of queues of fenced allocated ranges.
+ * @size: Size of the managed range.
+ * @align: Default alignment for the managed range.
+ */
+struct drm_suballoc_manager {
+	wait_queue_head_t wq;
+	struct list_head *hole;
+	struct list_head olist;
+	struct list_head flist[DRM_SUBALLOC_MAX_QUEUES];
+	size_t size;
+	size_t align;
+};
+
+/**
+ * struct drm_suballoc - Sub-allocated range
+ * @olist: List link for list of allocated ranges.
+ * @flist: List linkk for the manager fenced allocated ranges queues.
+ * @manager: The drm_suballoc_manager.
+ * @soffset: Start offset.
+ * @eoffset: End offset + 1 so that @eoffset - @soffset = size.
+ * @fence: The fence protecting the allocation.
+ */
+struct drm_suballoc {
+	struct list_head olist;
+	struct list_head flist;
+	struct drm_suballoc_manager *manager;
+	size_t soffset;
+	size_t eoffset;
+	struct dma_fence *fence;
+};
+
+void drm_suballoc_manager_init(struct drm_suballoc_manager *sa_manager,
+			       size_t size, size_t align);
+
+void drm_suballoc_manager_fini(struct drm_suballoc_manager *sa_manager);
+
+struct drm_suballoc *drm_suballoc_alloc(gfp_t gfp);
+
+int drm_suballoc_insert(struct drm_suballoc_manager *sa_manager,
+			struct drm_suballoc *sa, size_t size, bool intr,
+			size_t align);
+
+struct drm_suballoc *
+drm_suballoc_new(struct drm_suballoc_manager *sa_manager, size_t size,
+		 gfp_t gfp, bool intr, size_t align);
+
+void drm_suballoc_free(struct drm_suballoc *sa, struct dma_fence *fence);
+
+/**
+ * drm_suballoc_soffset - Range start.
+ * @sa: The struct drm_suballoc.
+ *
+ * Return: The start of the allocated range.
+ */
+static inline size_t drm_suballoc_soffset(struct drm_suballoc *sa)
+{
+	return sa->soffset;
+}
+
+/**
+ * drm_suballoc_eoffset - Range end.
+ * @sa: The struct drm_suballoc.
+ *
+ * Return: The end of the allocated range + 1.
+ */
+static inline size_t drm_suballoc_eoffset(struct drm_suballoc *sa)
+{
+	return sa->eoffset;
+}
+
+/**
+ * drm_suballoc_size - Range size.
+ * @sa: The struct drm_suballoc.
+ *
+ * Return: The size of the allocated range.
+ */
+static inline size_t drm_suballoc_size(struct drm_suballoc *sa)
+{
+	return sa->eoffset - sa->soffset;
+}
+
+#ifdef CONFIG_DEBUG_FS
+void drm_suballoc_dump_debug_info(struct drm_suballoc_manager *sa_manager,
+				  struct drm_printer *p,
+				  unsigned long long suballoc_base);
+#else
+static inline void
+drm_suballoc_dump_debug_info(struct drm_suballoc_manager *sa_manager,
+			     struct drm_printer *p,
+			     unsigned long long suballoc_base)
+{ }
+
+#endif
+
+#endif /* _DRM_SUBALLOC_H_ */